Our aim is to elucidate the molecular mechanisms of inducing tumor cell killing: the target-specific activation of one of the body's own cytotoxic mechanisms known as complement. We synthesized a hybrid protein consisting of the complement-activating cobra venom factor (CVF) and the monoclonal antibody to a tumor associated antigen. The nontoxic hybrid protein became a specific and efficient cytolytic agent for human melanoma cells in the presence of serum complement in vitro. We have devised a first in vivo model system (nude mice with intraperitoneal human melanomas) in which antibody-CVF conjugates show tumorsuppressive activity. The coupling of CVF offers advantages over coupling a toxin: (1)\absence of nonspecific toxicity; (2)\no need for internalization; (3)\complement exerts in vivo cytotoxicity through cell lysis and localized inflammation; and (4)\CVF might be replaced by a human C3b-derived molecule. This would result in a target-specific, nontoxic and nonimmunogenic cytotoxic agent. We investigate: (1)\the relationship between the uptake of individual complement components to cell death; (2)\the influence of the expression of the antigenic sites; and (3)\cellular defense mechanisms against complement attack, including their manipulation by metabolic inhibitors. Since no information is available on the mechanism of complement-mediated killing of nucleated cells initiated by a monoclonal antibody, parallel studies will be performed with a complement activating monoclonal antibody alone without CVF coupled to it. We have synthesized monoclonal antibody-CVF conjugates with three different heterobifunctional cross-linking reagents. The nature of the cross-linking reagent influences the rate of elimination and the interaction with plasma proteins of antibody-CVF hybrids. Some conjugates exhibited plasma half-times with up to 25 hrs, now allowing immunotherapeutic studies to be performed. It is our aim to produce a human C3-derived fragment that shows the same properties as CVF but is less immunogenic. During the last year, we could generate a 140,000 molecular weight C3-fragment by incubating human C3 with a specific protease in cobra venom. The new C3-fragment, termed C3o, shares some functional and structural characteristics with CVF. We will further characterize this fragment. As a tool for these studies, we will produce monoclonal antibodies to CVF to characterize functional sites of CVF and detect cross-reacting antigenic sites on human C3. (HI)